﻿968 Quantum Theory of Photographic Exposure. 



correction term in (14) is, for a given exposure, constant 

 throughout the array of classes, and v plotted against a 

 should give a straight line. If a were non-existent or 

 negligible, we should have a straight line for v plotted 

 against a itself. 



The aforesaid classification of targets and the corresponding 

 formula (14) will be used for analysing all the experiments 

 now in progress. For the present, we are able to quote only 

 one such set of results condensed in the following table. 



The targets (grains and clumps alike) were all divided into 

 five classes of equal breadth 2a = 0*60yit 2 , ranging from 0*20 

 to 0*80, from 0'80 to 1*40, etc., as shown in the first column, 

 which gives the average sizes a, in square microns. The 

 third column gives the observed number of targets surviving 

 for every N targets originally present, each of these data 

 being an average of counts on four different domains of the 

 plate. The fourth column contains the percentage number 



^ = 100^ of grains affected, as observed, and the fifth, as 



calculated by (14)> to wit, with 







72 = 0-255 



pei> 2 , 











<r = 0-008 V. 







a. 



N. 



N-Jc. 



y obs. 



ycaXc. 



A. 



0-50 



190-3 



1733 



8-9 



9-0 



-01 



110 



1400 



103-5 



26-0 



(20-6) 



(+5-4) 



1-70 



62-6 



43-0 



313 



31 -I 



+0-2 



2-30 



31-4 



18-7 



40-4 



40-3 



+0-1 



2-90 



19-8 



10-3 



48-0 



48-4 



-0-4 



The agreement is, apart from the second class, bracketed 

 as an "outlaw," almost perfect. The " cross-section " of 

 the darts, or what a may stand for, is agreeably about ten 

 times smaller, i. e. the diameter three times smaller than that 

 previously obtained with the same light source. This is not 

 to say that the reality and role of a is herewith settled. Yet 

 it is interesting that without cr, that is to say, with a in (14) 

 replaced by a itself, no choice of n yields such a close agree- 

 ment. As to the correction term due to the finite class 

 breadth, it may be mentioned that in the present ease it 

 amounts (as a subtrahendum from v) only to 0*030. 



Rochester, N.Y., 

 June 23, 1922. 



